B.Arivazhagan, Materials Technology Division, IGCAR, Kalpakkam, India, e-mail: chengaiari@yahoo.co.in M.Kamaraj, Metallurgical and Materials Engineering, IIT MADRAS, Chennai, India A study on influence of δ-ferrite phase on toughness of P91 steel welds P91 steel is also known as modified 9Cr-1Mo (P91) steel is widely used as a structural material in the construction of power plant components. In high-cr ferritic steels, toughness degradation in welds was caused by the presence of δ-ferrite phase in the martensite matrix. The δ-ferrite phase formation is influenced by factors such as chemical composition of welds, Creq and Nieq, heat input used during welding. As the δ-ferrite phase content increases there was reduction in toughness of welds below the specified requirement of 47 Joules as per the standard EN1557: 1997. The poor toughness of welds having δ-ferrite phase can be improved by prolonging the PWHT duration at 760 C. Flux system of consumables also influences the toughness of welds. Basic flux system produces welds having higher toughness than acidic flux system. This is due to microinclusion content of welds. The flux basicity, V and Nb content, and ferrite factor are interrelated and presented as a line diagram. The present study discusses about the role of chemical composition, and welding processes (SMAW and FCAW) on the formation of δ-ferrite phase in welds and its influence on toughness of welds. 1. Introduction example, fully tempered martensite in high-cr ferritic welds [8]. In high - Cr ferritic welds there was a Modified 9Cr-1Mo (P91) has been proposed as a steam significant reduction in toughness due to the presence generator material and also in other parts of nuclear of δ-ferrite phase in martensite matrix [8]. As the power plant components in the construction of Indian volume fraction of δ-ferrite increases above 2 %, there fast breeder programme [1]. The desirable properties considered for the above mentioned applications are high creep strength, resistance to thermal fatigue, and stress corrosion cracking (SCC) resistance [2]. Toughness of welds is an essential property required during hydro-testing of components fabricated from P91 steel by welding processes. Though high-cr steels are meant for hightable 1: Chemical composition of welds (wt. %) temperature applications, was a greater reduction in toughness of welds [8, 9]. In toughness is required in P91 steel welds to avoid brittle the present study, influence of δ-ferrite on toughness of fracture during hydro-testing of vessels. Various welds was discussed. Also the compositional factors (V specifications provide the requirement on toughness of and Nb content, and Ferrite factor), δ-ferrite content welds, for example, EN 1557:1999 [3]. As per the and welding processes (FCAW and SMAW) are above mentioned specification, impact energy of 47 interrelated by a two dimensional graph. Previously Joules (minimum average of 38 joules) is required at there has been no mention about such kind of graphical room temperature (20 C) for welds. Toughness is representation in existing literature about the factors influenced by variables such as welding process, influencing the toughness of P91 steel welds. composition of filler materials, post-weld heat treatment (PWHT) temperature and its duration [4, 5]. 2. Materials and Experimental work There are various methodologies available to enhance Standard 9Cr-1Mo steel (P9) in the normalized (960 C toughness in low alloy steels such as multipass welding for 0.3 h) and tempered condition (760 C for 1 h), and using thin layers [6], microinclusion content reduction modified 9Cr-1Mo steel (P91) plates in the normalized in welds [7] and homogenous microstructure, for 19
(1080 C for 1 h) and tempered (760 C for 2 h) condition were used for welding trials. The above mentioned base materials were cut into size of 220 x 100 x 12 mm, subsequently joined using processes such as flux-cored arc welding (FCAW), and shielded metal arc welding (SMAW) processes. The plates were butt welded (single V groove with included angle 60, root face of 1.6 mm, and root gap of 2.5 mm) in flat (1G) position. In the case of SMAW process, dissimilar (named as D1, D2, D3, and D4) and similar (S1, S2, and S3) weldments were prepared in order to assess the influence of chemical composition (V & Nb) and microstructure on toughness of welds. The study involved four base metal configurations, designated as D1, D2, D3, and D4, to produce different concentrations of Nb and V in weld metals and the details are as follows: D1: P9 (base metal) - P9 (filler) - P9 (base metal) D2: P9 (base metal) - P91 (filler) - P9 (base metal) D3: P91 (base metal) - P9 (filler) - P91 (base metal) D4: P91 (base metal) - P91 (filler) - P91 (base metal) Shielded metal arc (SMA) weld pads were prepared using 3.15 mm diameter electrode at a heat input of 1.64 kj / mm with preheat temperature of 200 C and interpass temperature of 250 C maintained during welding. Preheat and interpass temperatures were monitored using thermal crayons. SMA electrodes were baked in an oven at 250 C for 2 h before welding. Root layer was completed using GTAW process (filler of diameter 2.4 mm) to ensure good integrity. During root layer welding, pure argon (99.7 % purity) is used as a back shielding gas to avoid oxidation of weldments. The weldments were prepared in flat (1G) position. The flux-cored arc (FCA) welding was carried out using inverter based power source in flat (1G) position. The interpass and preheat temperatures were maintained as same in SMAW process. Two kinds of filler wires (diameter 1.2 mm) are used. One is acidic 20 type flux-cored wire and the other is basic type fluxcored wire. Acidic type flux-cored welds are designated as F1 and F2, whereas basic type flux-cored welds were designated as F3 and F4 respectively. The welding was carried out at a heat input of 1.4 kj /mm (Parameters: Voltage = 24-26V; Current = 80-190A; welding speed = 180 mm /min; flow rate of shielding gas = 22 liters / minute). There are two types of shielding gases were used namely 80 % Argon + 20 % CO2, and 95 % Argon + 5 % CO2 during welding. Defect-free weldments were subjected to post-weld heat treatment (PWHT) at 760 C for a period of 2 h after radiography inspection. Chemical composition analysis of base metal as well as welds were completed using optical emission spectroscopy (OES) technique. Nondestructive evaluation of weldment was carried out using radiography as per ASTM E-142 standard. Microstructural characterization was carried out using optical microscopy. To measure δ-ferrite phase content, quantitative metallography was completed using optical microscopy. Prior to microstructural characterization, samples were etched with Viella s solution (1 g picric acid, 5 ml HCl and 90 ml distilled water). Charpy Vnotch impact test was carried out as per ASTM E-23 at a test temperature of 20 C. 3. Results and Discussion 3.1 Chemical composition of base material and welds Base materials, P9 steel composition (wt. %) is as follows: C: 0.06; Mn: 0.35; Si: 0.27; Cr: 8.24 and Mo: 0.96; and P91 steel composition (wt. %) is as follows: C: 0.08; Mn: 0.39; Si: 0.50; Cr: 9.4; Mo: 1.0; V: 0.25; Nb: 0.09 and Ni: 0.13. Thermo-Calc Windows (TCW) software was used to determine the equilibrium transformation temperature based on composition of welds (Table 1) studied. The transformation point varies with chemical composition of welds as evident from the TCW Fig.1. TCW property diagram of weld S1 and F1
analysis. Ae1 being lower critical transformation point microinclusions are circular in shape as noticed from limits the PWHT temperature being used. Ae3 and Ae4 the microstructures. define the single phase region of austenite. In the case A typical SEM-EDS generated from a basic of SMA welds, silicon is higher (D3: 0.50 wt. %) in microinclusion is shown in Fig.3d. The chemical acidic flux-shielded welds than basic flux-shielded composition (wt.%) of a typical microinclusion is as welds (S3: 0.26 wt. %). Silicon being the strong ferrite follows: Oxygen = 13.80; Si = 11.40; Mn = 28; Fe = 35; stabilizer is less desirable in welds to enhance the Cr = 9.50; Ti =1.70.The microinclusions are of type toughness of welds. In basic flux-shielded metal arc SiO, MnO., etc. Figure.4 shows optical image of FCA welds, silicon content is low in welds. In FCAW, silicon weld2 showing δ-ferrite content in martensite matrix. In content is being influenced by flux composition as well as shielding gas composition. In basic fluxcored weld, F4, silicon is being higher due to inert shielding gas hence more amount of silicon (0.46 wt.%) transferred to the resultant weld. Higher the silicon content, more will be the volume fraction of δ-ferrite phase in weld. The δ-ferrite morphology also varies depending on the silicon content. With increase in silicon content, δ-ferrite phase is observed at grain boundary as well as within the grains. The TCW windows generated property diagram for a shielded metal arc weld ( S1 ) is shown in Fig.1a. Similarly for a FCA weld ( F1 ), TCW Fig.2. P91 steel in normalized and tempered condition TEM image TEM-EDS analysis generated property diagram is shown in Fig. 1b. The observed phases were similar in both the present study, basic FCA weld shows higher δcases except the transformation points of welds. ferrite phase than SMA weld. FCA weld shows highest with few precipitates observed within the δ3.2 Microstructure analysis of welds and δ-ferrite, ferrite phase. Another reason for higher amount of δbase metal ferrite phase in weld F4 is it has higher amount of P91 steel microstructure in the normalized (1080 C-1 silicon content than other welds. In the SMA welds h) and tempered (760 C- 2 h) is shown in Fig.2a. Well studied, there was absence of precipitates within the δdefined sub-grain structure is formed after normalizing ferrite phase. and tempering treatment. A blocky, near circular Figure 5 shows the TEM image of microstructure Vs shaped precipitate within the lath is observed. TEM- toughness of acidic flux-cored arc welds (Fig 5a: 760 C EDS analysis performed on the precipitate (as shown by 2h, and Fig.5b: 760 C 5 h conditions). These arrow in Fig.2a) reveals the following composition. V = images highlight the microstructural changes occurring 1.3 %, Cr = 3.5 %, Fe = 31.9 % and Nb = 63.4 %. From in the prolonged post-weld heat treatment of P91 steel the composition, the precipitate is of carbonitride (MX welds. As the PWHT duration increases there was an type) precipitate. increase in size of the precipitates and sub-grain The acidic flux-cored weld, Fig.3a, shows numerous formation in the weld. Similar kind of microstructural microinclusions with varying sizes (fine: 2-5 µm and changes were observed in SMA welds. It was coarse: > 5µm). This has been due to more amount of commonly observed that as-weld microstructure shows oxygen (692 ppm) generated during welding and needle like structure hence the toughness was low. After trapped during solidification as oxide microinclusions. PWHT at 760 C 2 h, there was change in morphology SEM-EDS analysis shows the presence of elements in of precipitates. As the PWHT duration increases, there microinclusion associated with nature of flux used was an increase in quantity of precipitates observed. during welding. From the inclusion analysis (Fig.3b), This phenomenon is commonly observed in the welds strong peaks of Ti, Al and Si were observed. Its (FCA and SMA welds) studied. chemical composition (wt. %) details are as follows: Al 3.3 Influence of flux basicity, vanadium and = 4.93; Si = 5.29; P = 0.25; S = 0.46; Mn = 19.55; and Ti =32.88 suggesting the presence of Al2O3, TiO2 and niobium and Ferrite Factor (FF) on MnS etc. In basic flux-cored arc welds, Fig.3c, the toughness of welds. microinclusions are fine, microscopic in size. They are widely spaced as evident from the fig.3c. By 3.3.1 Ferrite Factor (FF) Vs (V + Nb) comparison of figures (3a and 3c), It has been content of welds understood that basic flux-cored welds have less Schneider formula has been widely used to predict weld microinclusion content (Size: 2-5 µm: 50 numbers, and metal microstructure in high alloy steels. It has more > 5 µm: Nil) than acidic flux-cored welds (Size: 2-5 weightage for elements such as nitrogen, molybdenum µm: 160 numbers, and > 5 µm: 6 numbers). All weld and niobium. For prediction of δ-ferrite phase 21
formation, Schneider s formula is now widely used [8]. Hence in the present study, Schneider formula is used to evaluate Creq (Eq.1), and Nieq (Eq.2) based on weld metal composition. The Schneider formula is as follows: Creq = Cr + 2Si + 1.5Mo + 5V + 1.75Nb + 0.75W (Eq.1) Nieq = Ni + 0.5Mn + 30C + 25N + 0.3Cu (Eq.2) To obtain weld metals with fully martensitic microstructure, it appears necessary to consider both a Schneider chromium equivalent (Eq.1) value of lower and FF, the V and Nb content along with flux system must be considered to get a good tough weld. Fig. 6a and Fig.6b show the effect of flux basicity on weld metal toughness. These figures show to develop the required toughness of 47 Joules by PWHT at 760 C 2 h in acidic flux (Fig.6a) system, V + Nb content should be less than 0.18 %, whereas basic flux system (Fig.6b) can have higher V + Nb content of 0.31 %. This is due to higher amount of microinclusions in acidic flux system that can cause severe reduction in toughness of welds. In acidic flux systems, weld microinclusions are coarser in size (> 5 µm) as compared with basic flux systems (< 5 µm). The coarse microinclusions act like brittle notches that degrade toughness as reported elsewhere [5]. Due to combined influence of microinclusion size and V+Nb content, acidic flux system can tolerate lesser V + Nb content (0.18 wt.%) than basic flux system (0.31 wt.%). Another noteworthy feature observed from the Fig.6a and Fig.6b is that V+Nb content has more significant influence than ferrite factor and also flux system to meet the specified 47 joules. 3.4. Charpy V-notch impact toughness of welds at room temperature (20 C) Toughness evaluation was done using full size Charpy V-notch specimens at room temperature. The reported value is the average of three values. Table 2 lists the impact energy of (c) (d) various welds at room temperature Fig.3. SEM image of microinclusion content in acidic weld SEM-EDS (20 C). From the data, it has been analysis of acidic weld (c) basic weld (d) SEM-EDS analysis of basic weld inferred that weld having the lowest V + Nb content (0.06 %) content with less δ-ferrite phase exhibits than 13.5, and the difference between the chromium maximum toughness of 65 Joules among the welds (Creq ) and nickel (Nieq) equivalents known as ferrite studied. In acidic flux shielded metal arc welds, despite factor (FF) to be lower than a value of 8. From the the high microinclusion content, toughness can be chemical composition of welds, it was inferred that improved by having low V + Nb content. The welds even though welds may have relatively high austenite having forming elements (High Nieq) and consequently toughness less relatively low FF. They may not develop the required than 47 Joules toughness with PWHT of 2 hours. These welds may can be further contain high V and Nb, which will reduce toughness. It improved by can be said that higher strength and lower ductility are prolonging the obtained with increase of V contents. The effect of Nb post-weld heat content on material properties is the same as that of treatment from vanadium [9], but the effects is almost invisible over 2 hours to 5 0.01 mass %. The impact properties degrade with the hours. It has increase of vanadium content. In contrast, the influence been observed of Nb content on the degradation of impact properties from the data in Fig.4. Optical, as-weld tends to saturate over 0.01 mass %. These welds table 2, welds microstructure showing the presence especially in acidic flux system, require more hours of having basic of δ-ferrite in martensite matrix PWHT. This in addition to the knowledge of Creq, Nieq flux systems 22
the weld D3 having highest ferrite factor (9.64), have lowest transformation temperature to austenite (1065 C) hence δ-ferrite phase is observed in weld D3. On the other hand, weld S3 having lowest ferrite factor (6.11) have the highest transformation temperature (1140 C) hence δ-ferrite phase is completely absent in the weld. In FCA welds, the weld having highest ferrite factor (weld F4 ; FF: 8.31; 1070 C) δ-ferrite phase is observed. On the other hand, the weld having Fig.5. Influence of PWHT duration on microstructure of welds 760 C- 2h the lowest ferrite factor (weld F1 ; 760 C- 5h FF: 7.07; 1089 C) shows complete absence of δ-ferrite phase. exhibit better toughness than acidic flux systems. In the case of acidic welds, after prolonged post-weld heat 4. Conclusion treatment there was less increase in toughness than basic welds. This is due to dominant influence of flux 1. It is possible by increasing the PWHT duration from system on toughness of welds. The low toughness of 2 h to 5 h, welds having poor toughness (less than 47 weld F4 has been attributed to the following factors: Joules) along with high FF can also meet the minimum High ferrite factor (8.31), Creq (13.30), more amount of requirement of 47 Joules. vanadium and niobium (0.32 wt. %), highest amount of 2. Higher amounts of vanadium, niobium and silicon silicon (0.46 wt. %). The above mentioned factors are content in welds are possible by having basic flux Fig.6. Effect ofv and Nb content on toughness of welds after PWHT at 760 C 2h acidic weld basic weld explained in detail as follows. With more amount of vanadium and niobium there will be an enhancement in precipitation of carbonitrides with consequent reduction in toughness [10]. High silicon promotes more amount of δ - ferrite since silicon is being a strong ferrite stabilizer in P91 steel [4]. In weld with a high FF, δ-ferrite is stable over a wide temperature range. Consequently, while δ-ferrite forms during heating, the amount possibly increasing with increasing arc energy, the reverse transformation to austenite takes place at relatively low temperature and hence it is retarded [11]. As a result, retained δ-ferrite phase levels may tend to increase with increasing heat input in welds having high FF. On the other hand, low FF material produces less δ-ferrite phase at peak temperatures, but the reverse transformation to austenite occurs at higher temperatures, and the net result could be a reduction in the final δ-ferrite phase content [11]. In the case of shielded metal arc welds, system in consumables. In essence, basic flux systems can be preferred in which higher amounts ofv, Nb and Si in weld can be tolerated and the required toughness of 47 Joules can be obtained. Also to a certain extent the δ-ferrite (up to 2 %) can be present in basic fluxshielded welds without impairing the toughness. 3. Nieq and Creq diagrams and limiting ferrite factor alone may not be sufficient to define good welds. The graph connecting ferrite factor (FF), and V+Nb content can give the zone where the required toughness of 47 Joules can be obtained with PWHT of 2 hours, especially in the case of acidic flux system. 5. References 1. M.Sireesha, S.K. Albert, and S.Sundaresan, Microstructure and mechanical properties of weld fusion zones in modified 9Cr-1Mo steel. Journal of Materials Engineering Performance, 10 (3), 2001, 320-330. 23
2. A.M.Barnes, The effect of composition on microstructural development and mechanical properties of modified 9%Cr-1%Mo weld metals. In: Strang, A., Cawley J., Greenwood, G.W. (Eds.), Proceedings on Microstructural Stability of Creep Resistant Alloys for High Temperature plant applications. The Institute of Materials London, 1998, 339-360. 3. E.L.Bergquist, Consumables and welding modified 9Cr-1Mo steel. Svetsaren, (1-2), 1999, 2-5. 4. S.Dittrich, V. Gross, and H.Heuser, Optimized and advanced P (T) 91 filler metals. In: Proceedings of National Welding Seminar India, 1994, 4R1-4R5. 5. B. Arivazhagan, S.Sundaresan and M.Kamaraj, Effect of TIG Arc Surface Melting Process on weld metal toughness of modified 9Cr-1Mo (P91 ) steel, Materials Letters, 62 (17-18), 2008, 2817 2820. 6. Z. Zhang, A.W. Marshall, and G.B. Holloway, FluxCored Arc Welding: High Productivity Welding Process for P91 Steels. EPRI Conference on 9Cr Materials Fabrication and Joining Technology, USA, 2001, 1-14. 7. M.L.Hu, T.C.Yang, H.C.Wu and C.Chen, Fusion zone purification of metal containing oxide inclusions via laser welding, Journal of Materials Science 40 (2005) 4125-4127. 8. J.Onoro, Martensite microstructure of 9-12% Cr steel weld metals, Journal of Materials Processing Technology, 180, 2006, 137-142. 9. A. Barnes, The Influence of Composition on microstructural development and toughness of modified 9%Cr-1 Mo weld metals, TWI report, 509 / 1995, 1995. 10. Takashi Onizawa, Takashi Wakai, Masanori Ando and Kazumi Aoto, Effect of V and Nb on precipitation behavior and mechanical properties of high Cr steel, Nuclear Engineering and Design, 238 (2008) 408-416. 11. Roger Panton-Kent, Phase balance in 9%Cr1%Mo steel welds, TWI Bulletin, January / February 1991. Table 2 Influence of FF, flux system and V+Nb contents on toughness of P91 steel welds 24